On 2008-11-08 at 09:16:07 [+0100], Humdinger <humdingerb@xxxxxxxxxxxxxx> wrote: > Axel Dörfler wrote: > > They are actually named, see: > > http://dev.haiku-os.org/browser/haiku/trunk/headers/private/kernel/boot_splash.h > > > > When BOOT_SPLASH_STAGE_1_INIT_MODULES lits up (the first one), the > > kernel core services are already fully initialized and up and running. > > If you need explanations somewhere, please ask. > > OK, coming back to this, would you say the following is accurate? Apart > from stage 2 > "BOOTSTRAP_FS", where "loading kernel is probably wrong"... > > STAGE_1_INIT_MODULES = 0, > Atom - Initializing modules. > > STAGE_2_BOOTSTRAP_FS, > Disk with magnifying glass - Loading kernel. Nope, the kernel has already been loaded before stage 1. In this phase the root file system is created ("/"), the devfs mounted ("/dev"), and a few symlinks in the rootfs are created. The devices entries itself aren't available yet. IMHO this activity doesn't really deserve a separate icon, since it's only a matter of a few microseconds. > STAGE_3_INIT_DEVICES, > Plug-in card - Initializing devices. Initializes the device manager. The devices themselves are mostly lazily initialized when first accessed. > STAGE_4_MOUNT_BOOT_FS, > Boot disk - Mounting boot disk. > > STAGE_5_INIT_CPU_MODULES, > Chip - Initializing CPUs. This just loads the CPU module responsible for this CPU brand/family/model, which allows access to specific features (like setting the MTRR (memory type range registers)) and provides optimized versions of certain functions (or would do that, if someone wrote them). > STAGE_6_INIT_VM_MODULES, > Folder - Initializing virtual memory. Not really -- the VM is up before stage one and swap support is initialized in phase four (as soon as the swap file can be accessed). Various subsystems do final initializations that potentially require modules to be loaded from the boot disk. CU, Ingo